{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/online-learning-for-effort-reduction-in","title":"Online Learning for Effort Reduction in Interactive Neural Machine Translation","arxiv_id":"1802.03594","date":"2018-02-10","proceeding":null,"authors":["Álvaro Peris","Francisco Casacuberta"],"abstract":"Neural machine translation systems require large amounts of training data and\nresources. Even with this, the quality of the translations may be insufficient\nfor some users or domains. In such cases, the output of the system must be\nrevised by a human agent. This can be done in a post-editing stage or following\nan interactive machine translation protocol.\n  We explore the incremental update of neural machine translation systems\nduring the post-editing or interactive translation processes. Such\nmodifications aim to incorporate the new knowledge, from the edited sentences,\ninto the translation system. Updates to the model are performed on-the-fly, as\nsentences are corrected, via online learning techniques. In addition, we\nimplement a novel interactive, adaptive system, able to react to\nsingle-character interactions. This system greatly reduces the human effort\nrequired for obtaining high-quality translations.\n  In order to stress our proposals, we conduct exhaustive experiments varying\nthe amount and type of data available for training. Results show that online\nlearning effectively achieves the objective of reducing the human effort\nrequired during the post-editing or the interactive machine translation stages.\nMoreover, these adaptive systems also perform well in scenarios with scarce\nresources. We show that a neural machine translation system can be rapidly\nadapted to a specific domain, exclusively by means of online learning\ntechniques.","url_abs":"http://arxiv.org/abs/1802.03594v2","url_pdf":"http://arxiv.org/pdf/1802.03594v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"online-learning-for-effort-reduction-in","repo_url":"https://github.com/lvapeab/nmt-keras","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.03594","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}